PSI - Issue 16

Borys Paton et al. / Procedia Structural Integrity 16 (2019) 176–183 Borys Paton et al. / Structural Integrity Procedia 00 (2019) 000 – 000

180

5

monitoring. Repair, mounting, welding and other kinds of operations conducted near the objects of control, can be additional sources of interference. To solve these problems, special software was developed, allowing maximum lowering of background noise influence on operation of basic algorithm used during monitoring. Testing began during the cold season, at working parameters of pipeline operation and was continued after their unloading. Preliminary data already showed that the method is highly informative, but that there are also certain problems in system setting up under the conditions of high level of technological noise. Testing and setting of instrumentation modes were first performed. These are the modes that directly influence information acquisition and cannot be changed during computer reprocessing of obtained data. First of all operating modes with different gain factors were tested, and it was established that optimum AE data were obtained at amplifier operation in the logarithmic mode, in which they are pre-set to operate. The following tests dealt with the level and type of thresholds of amplitude discrimination. Considering the data of initial testing and high noise level of the object of control proper, it was found that optimum threshold settings are floating ones, i.e. the threshold changes automatically with the total noise level, recorded by the particular transducer. Optimum floating threshold should be slightly higher than the current noise level at the transducer. Threshold value is corrected every 30 s. Such settings ensured the required measurement sensitivity, but the incoming information flow, allowing for strong noises, was too high. This created the need for additional filtering of the data, which was achieved by doing additional tests on the object and analyzing the data of testing samples from analog materials at high temperatures. In particular, effectiveness of filtering by such AE signal parameters as amplitude and rise time was considered. Preliminary results of filtering application showed the following:  application of filtering by amplitude leads to lowering of the number of AE events, fitting into the formed clusters. Warning levels are lowered, accordingly  at application of filtering by amplitude part of AE events, corresponding to real AE sources in pipeline material, are rejected  at raising lower boundary of filter by amplitude above 60 – 65 dB, AE events are rejected almost completely, at its setting below 55 dB the filter becomes ineffective  filter by “rise time” parameter can be effectiv ely used in monitoring settings  filtering by this parameter leads to lowering of warning level; here active AE sources form clusters and can be clearly singled out from general background noise  possibility and effectiveness of filtering by duration, frequency, velocity of AE signals was analyzed in a similar way, and optimum filtering bands were selected, which provide sufficient information for searching for hazardous locations and generation of appropriate warming. As after transition to full-time operation it is intended to use breaking load prediction by AE data, the obtained data were used to verify prediction algorithms used in the system. Here, prediction and warming about the danger are conducted in keeping with the given tables. Calculation of predicted breaking load is performed automatically.

Personnel actions in keeping with danger indicator readings for continuous monitoring are specified at the Table 1.

Table 1. Personnel actions at different readings of danger indicator. Warming level Indicator color Personnel actions 1 Green

Standard mode. Continue operation

2

Yellow

Attention. At appearance of predicted breaking pressure level and its exceeding the working pressure by more than two times – continue operation Assess predicted breaking pressure level by indicator readings. At predicted level exceeding the working level less than two times, pay greater attention to readings. At predicted pressure decreasing to 1.3 of working pressure – stop operation. Perform additional check of problem component.

3

Orange

4 Stop operation. After appearance of intermittent sound signal, immediately lower the load Remark: Table 1 data were obtained for instrumentation with EMA 3.9 parameters. Testing of the mounted system was conducted in standard mode of continuous monitoring and in the form of one-time tests. Standard mode envisages automatic saving of files of 30 min duration with subsequent start of new file recording. Red, pulsing